On Sun, Dec 4, 2016 at 11:11 PM, Andreas Karlsson <andr...@proxel.se> wrote:
> On 12/04/2016 02:12 PM, Michael Paquier wrote:
>> One last thing that I think is missing in this patch is for users the
>> possibility to check via SQL if the SSL context is actually loaded or
>> not. As the context is reloaded after all the new values are
>> available, with the current patch users may see that ssl is set to on
>> but no context is loaded. So why not adding for example a read-only
>> parameter that maps with SSLLoaded?
> The other three issues are easy to fix, but this one is a bit trickier. Do
> you mean that we should add another GUC here which is read-only?

Yes, that's what I meant. It is hard to track if the reloading has
been effective or not.

> Does this have a precedent in the code?

data_checksums in guc.c is an example, it is marked with
GUC_NOT_IN_SAMPLE | GUC_DISALLOW_IN_FILE and its value is updated with
SetConfigOption() when the control file is read. The idea would be to
do something like that with LoadedSSL.

Sent via pgsql-hackers mailing list (pgsql-hackers@postgresql.org)
To make changes to your subscription:

Reply via email to